ABSTRACT
BACKGROUND: Reliable identification of endothelial cells is a prerequisite for understanding vascularity changes in many cardiovascular diseases and therapeutic interventions. With the rising use of mouse models of disease and genetic manipulation, a consistent system to label murine endothelial cells in normal and diseased tissues would be an invaluable tool. METHODS: Immunohistologic and histochemical methods were used to label endothelial cells in normal and infarcted heart as well as small intestine. Isolectin B(4) or antibodies to thrombomodulin, vWF, Tie-2, VE-cadherin, CD34, and CD31 were used to immunostain tissues fixed in either 4% formaldehyde (prepared fresh from powdered paraformaldehyde, PF), methyl Carnoy's (MC), zinc (Zn) (all paraffin-embedded), or frozen sections. In addition, we used beta-galactosidase histochemistry in frozen sections from the Tie-2/beta-galactosidase transgenic mouse, in which the lacZ reporter gene is driven by the endothelial-specific Tie-2 promoter. RESULTS AND CONCLUSIONS: Immunohistologic localization of endothelial cells was best accomplished using anti-CD31 in Zn-fixed, paraffin-embedded tissues. Antithrombomodulin staining also worked in Zn-fixed tissues as well as frozen tissues, but additional steps were required to reduce background. The beta-galactosidase reporter system also marked endothelial cells in frozen Tie-2 transgenic mouse tissues; however, this system required careful control of fixation and optimization of histochemistry and was not amenable to double labeling. Lectin staining gave complete labeling of endothelial cells but cross-reacted intensely with macrophages in the infarcted heart. We conclude that anti-CD31 immunostaining in Zn-fixed, paraffin-embedded murine tissue offered superior morphology and permitted optimal identification of proliferating endothelial cells during infarct repair.
